AI平台工程师
AI Platform Engineer
#### Nextdata 正在寻找一位经验丰富的 AI 平台工程师，参与一个为期三个月的初始项目，专注于构建和验证一个具备 AI 能力的数据平台功能。
#### 这是一个需要亲自参与实施的角色。你将快速熟悉现有项目，并构建后端服务、代理工具、API 和数据集成。该职位不涉及面向客户的职责，相较于高级职位，对架构的关注较少。
#### 如果项目成功，可能会继续并发展为更广泛的 Nextdata 产品。
#### **你将负责**
- 快速理解现有代码库、需求和技术方向。
- 构建用于受控数据访问的后端服务、AI 代理、工具和 API。
- 实现兼容 MCP 的端点或类似的代理接口。
- 集成 SQL、文档、API、元数据、语义模型和向量搜索。
- 构建检索、工具选择、上下文构建和基于事实的响应流程。
- 添加访问控制、策略执行、审计日志、测试和可观测性。
- 在三个月的项目范围内交付一个可靠且有文档的系统。
#### **我们寻找的人选**
- 在后端系统、数据平台或分布式系统方面有丰富经验。
- 有使用 LangChain、LangGraph 或类似框架构建代理式 AI 应用的实际经验。
- 有 RAG、检索、语义搜索、工具调用和多步骤代理工作流的经验。
- 强大的 Python 和 API 开发技能。
- 有 SQL、文档、元数据系统和向量搜索的经验。
- 熟悉 MCP 或类似的代理接口。
- 理解访问控制、策略、数据血缘、数据质量、PII 保护和可审计性。
- 能够在不熟悉的代码库中快速上手。
- 对范围、权衡和生产就绪程度有良好的判断力。
### **加分项**
- 有数据产品、数据网格、语义模型、目录或治理平台的经验。
- 有 MCP 服务器、工具注册表或多步骤代理的经验。
- 有 Databricks、Snowflake、BigQuery、Spark、DuckDB、Postgres、图数据库或向量数据库的经验。
- 熟悉 OAuth、OIDC、SAML、SSO、RBAC、ABAC、SCIM 或策略引擎。
- 有评估检索质量、工具准确性、基于事实性和故障模式的经验。
### **合同范围**
- 初始期限：三到四个月。
- 重点：亲自参与实施和交付。
- 不涉及面向客户的职责。
- 可能延长
查看英文原文
#### Nextdata is looking for an experienced AI Platform Engineer for an initial three-month project focused on building and validating an AI-enabled data platform capability.
#### This is a hands-on implementation role. You will quickly ramp up on an existing project and build backend services, agent tools, APIs, and data integrations. The role has no customer-facing responsibilities and less emphasis on architecture than a principal-level position.
#### If successful, the project may continue and develop into a broader Nextdata product.
#### **What You’ll Do**
- Quickly understand the existing codebase, requirements, and technical direction.
- Build backend services, AI agents, tools, and APIs for governed data access.
- Implement MCP-compatible endpoints or similar agent interfaces.
- Integrate SQL, documents, APIs, metadata, semantic models, and vector search.
- Build retrieval, tool selection, context construction, and grounded response flows.
- Add access controls, policy enforcement, audit logging, tests, and observability.
- Deliver a reliable, documented system within the three-month project scope.
#### **What We’re Looking For**
- Strong experience with backend systems, data platforms, or distributed systems.
- Practical experience building agentic AI applications using LangChain, LangGraph, or similar frameworks.
- Experience with RAG, retrieval, semantic search, tool calling, and multi-step agent workflows.
- Strong Python and API development skills.
- Experience with SQL, documents, metadata systems, and vector search.
- Familiarity with MCP or similar agent interfaces.
- Understanding of access control, policies, lineage, data quality, PII protection, and auditability.
- Ability to become productive quickly in an unfamiliar codebase.
- Good judgment on scope, trade-offs, and production readiness.
### **Nice to Have**
- Experience with data products, data mesh, semantic models, catalogs, or governance platforms.
- Experience with MCP servers, tool registries, or multi-step agents.
- Experience with Databricks, Snowflake, BigQuery, Spark, DuckDB, Postgres, graph databases, or vector databases.
- Familiarity with OAuth, OIDC, SAML, SSO, RBAC, ABAC, SCIM, or policy engines.
- Experience evaluating retrieval quality, tool accuracy, groundedness, and failure modes.
### **Contract Scope**
- Initial term: three to four months.
- Focus: hands-on implementation and delivery.
- No customer-facing responsibilities.
- Potential extension if the project is successful.